This talk will discuss how robots can be improved by integrating Large Language Models (LLMs) as a source of common-sense knowledge. Traditional knowledge-based systems are limited, but LLMs provide a more dynamic and intuitive approach. They allow robots to interpret and respond to complex human environments more naturally, providing a comprehensive understanding of the context of their actions. This is particularly important in community-centric applications where understanding social norms and behaviors is essential.

     The talk will focus on integrating LLMs into the robots' state machines, which allows for a seamless blend of robotic programming with advanced language understanding capabilities. This enhances the robots' ability to interact and operate within human-centric environments. Technical aspects of this integration will be discussed, including adapting LLMs to understand the specific action capabilities of robots, as well as the constraints and forbidden actions pertinent to various contexts.

     The presentation will also cover the potential challenges and ethical considerations of implementing LLMs in robotics, especially in community settings. This includes addressing concerns about privacy, safety, and the responsible use of AI in public spaces.

     Overall, this talk aims to show the transformative potential of LLMs in revolutionizing service robotics. Attendees can look forward to gaining valuable insights into the future of robot-human interaction and the evolving role of AI in this dynamic field.

In this talk, we propose a transformer-based semantic robot with a computational intelligence (CI) mechanism designed for use in an educational co-learning environment, where teachers, teaching assistants, and students interact with the CI robot and attention ontology to enhance the learning process. The approach is applied in two distinct applications. The first, focusing on student-machine co-learning with writing performance evaluation, involves an attention-based mechanism for curating learning content from students, which is further refined by a preprocessing mechanism with expert-based fuzzy numbers. The second, concentrating on student-machine co-learning with speaking performance evaluation, introduces a Meta AI Universal Speech Translation (UST) mechanism that translates content into English and Taiwanese speeches, as well as into English and Chinese texts. This transformer-based robot for computing semantic similarities employs a trained semantic SBERT model to analyze student-machine co-learning contents. Given the large size of the co-learning content with the ontology model, we implement a chunk-based approach for processing. This method enables effective comparison of the extensive student-provided learning content with the evaluative content from teachers and teaching assistants. Additionally, a human intelligence-based robot, equipped with a CI assessment mechanism based on fuzzy numbers, evaluates performance and adjusts the evaluation content of teachers and teaching assistants based on human intelligence (HI) fuzzy numbers. Experimental results indicate that the proposed CI robot can reduce teachers’ burden and objectively evaluate student-machine co-learning performance, thereby narrowing the gap in actual student-machine co-learning performance. Furthermore, it aids in assessing student-machine co-learning performance and understanding, creating a more personalized and effective learning environment. 

Agriculture is a vital industry because of its contribution to the stability of people's lives, income and national employment as well as providing raw materials for other industries. Then the agricultural sector has a direct impact on all segments of society, where in Indonesia within last 10 years the number of farmers has decreased while the population has increased. This causes the farmer workforce to fall even further due to the weight and amount of work. This is exacerbated by the fact that the harvest results are not optimal, especially because the parameters that cause the success or failure of the harvest are not measured.

The latest technological developments that have covered various fields provide solutions for the agricultural sector. Unmanned Aerial Vehicle (UAV) technology has great potential for mapping land and the location of robots, while Unmanned Ground Vehicles (UGV) can be tasked with carrying out work directly on agricultural land automatically. Therefore, in this research, smart-farming through the development of air and land robots is being developed to accelerate agricultural technology towards precision agriculture, maximizing crop yields. First, the algorithm on the computer will produce land mapping and process the data to obtain the required path. This route is sent to the UGV on agricultural land as a map of the UGV's movement. Next, a UAV equipped was developed to automatically map plant areas, using photogrammetry technology. We can carry out regular mapping with aerial robots and then get information such as plant health, monitoring, fertilizer nutrient analysis, area analysis and harvest result prediction without going directly to the land.

The implementation of the robot technology will be able to develop and maximize agricultural yields through precision agriculture.

In today's data-driven landscape, optimizing decision-making is paramount. This talk delves into the critical roles of anomaly detection and recommender systems in modern analytics. It begins by highlighting the significance of anomaly detection within the realm of intelligent data streams and its impact on decision-making processes. Streaming data frequently arrives in massive quantities and rapidly, making manual inspection and analysis difficult and time-consuming. Detecting anomalies in real-time streams is challenging because the detector must process data and make decisions in real-time. The talk explores the utilization of machine learning approaches for anomaly detection and provides real-world examples from intelligent data streams. 

    Moving on to recommender systems, the talk delves into personalization, emphasizing its effect on enhancing user experiences, explaining the inner workings of recommendation engines, exploring advanced techniques for tailoring recommendations, and presenting case studies. Finally, the talk will demonstrate how anomaly detection is a valuable addition to recommender systems, how they synergize to significantly enhance the decision-making process, and present real-world applications of this fusion.

In healthcare, technology can help or hinder.  Healthcare workflows can be hindered by complex systems, interoperability issues and a zoo of well-meaning applications which ultimately see low end user adoption.  To fix this, we must design technology that meets user needs and is simple to use. I will talk about current projects addressing issues in remote patient care.  These projects focus on developing technology designs that are easy for users to use, cost-effective, and scalable for health care providers.  The goal is to make technology work better in healthcare by keeping it straightforward and practical.  The projects focus on patients with chronic disease who are in need to regular home monitoring to minimise their risk of hospital admission, which includes monitoring of vital signs for Chronic Obstructive Pulmonary Disease (COPD), Congestive Heart Failure (CHF) and adherence to complex medication plans.